f.cam6_3_119.FLTHIST_ne30.r328.opt.002

[cecilehannay]

Description: Same as #345 +

se_hypervis_subcycle_sponge = 2->4
se_nu_top = 5E5->1E6

@PeterHjortLauritzen ram two year in that configuration and it was stable.

Detailed description The purpose of this run is to test various optimizations to the dynamical core for increased throughput (see details below). Only tested in 2 year long run so we will see if it is stable for a multi-year run ...

Same as https://github.com/NCAR/amwg_dev/issues/344 with SourceMods

/glade/u/home/pel/src/for-science-opt/dimensions_mod.F90
/glade/u/home/pel/src/for-science-opt/global_norms_mod.F90
/glade/u/home/pel/src/for-science-opt/prim_advance_mod.F90
/glade/u/home/pel/src/for-science-opt/prim_advection_mod.F90
/glade/u/home/pel/src/for-science-opt/prim_driver_mod.F90
/glade/u/home/pel/src/for-science-opt/viscosity_mod.F90

and namelist additions:

se_nsplit = 1
se_rsplit = 6
se_sponge_del4_nu_div_fac  = 1
se_hypervis_subcycle = 1

se_nu_div = 1.0E15
se_nu     = 1.0E15
se_hypervis_subcycle_sponge = 2->4
se_nu_top = 5E5->1E6

Case directory: Locally (if still available): /glade/p/cesmdata/cseg/runs/cesm2_0/f.cam6_3_119.FLTHIST_ne30.r328.opt.002

On github: https://github.com/NCAR/amwg_dev/tree/f.cam6_3_119.FLTHIST_ne30.r328.opt.002

Sandbox: Locally (if still available): /glade/work/hannay/cesm_tags/cam6_3_119

On github: (https://github.com/ESCOMP/CAM/tree/cam6_3_119)

Diagnostics: ADF diags (if available) https://webext.cgd.ucar.edu/FLTHIST/f.cam6_3_119.FLTHIST_ne30.r328.opt.002/atm/

Contacts:

@cecilehannay @PeterHjortLauritzen @adamrher @JulioTBacmeister

[PeterHjortLauritzen]

Could we also get diags against OBS? Thanks

[PeterHjortLauritzen]

@adamrher was right that increased sponge layer diffusion has a very detrimental effect on U (below is plot against OBS):

Plan: reduce del2 damping in sponge which will require to increase del4 (less detrimental than del2) but the model won't run as fast ... will look at that once higher priority stuff is done ...